In contrast to the "traditional" methods of multispectral remote sensing techniques, where only a limited number of spectral bands were available for interpretation, the modern hyperspectral remote sensing methods can utilise the data from spectrometers recording more than 500 bands simultaneously.

Through this method, a broad range of the electromagnetic spectrum, mostly between 0.4 μm (visible) and 2.5 μm (infrared) is divided in small intervals (spectral sampling) and recorded in one band for every single pixel.

When light interacts with a material on the surface, the light of a certain wavelength is absorbed, while at other wavelengths it is reflected. Every material has a very specific reflectance value at a given wavelength. When the continuous reflectance spectrum of a sample is measured, information can be obtained about the composition of the material (like a fingerprint or DNA information).

Reflectance spectra represented in the different bands acquired by airborne hyperspectral sensors can - after calibration and atmospheric correction - be directly compared to field/laboratory reflectance spectra (spectral libraries), thereby enabling the determination of materials on the ground, without the need for pre-existing and extensive ground reference.

Further Information:

- Priciples of hyperspectral imaging: Hyperspectral_Doc1.pdf, 1MB

- Pre-processing of hyperspectral data acquired: Hyperspectral_Doc2.pdf, 3,4MB

- Application Examples: Hyperspectral_Doc3.pdf, 1MB